Downloads · 30 days
68
10% of all-time downloads
RichardErkhov/sarvamai_-_sarvam-2b-v0.5-gguf
sarvamai_-_sarvam-2b-v0.5-gguf is a machine learning model from RichardErkhov. Use it for the machine learning task on the model card, and read the license before you ship it in a product.
Downloads · 30 days
68
10% of all-time downloads
All-time downloads
674
Public
Repo size
28.8 GB
Likes
0
Public
Click a slice to open those files.
.gguf33.4 GB · 100%
From the Hugging Face model README
Quantization made by Richard Erkhov.
sarvam-2b-v0.5 - GGUF
| Name | Quant method | Size |
|---|---|---|
| sarvam-2b-v0.5.Q2_K.gguf | Q2_K | 0.91GB |
| sarvam-2b-v0.5.IQ3_XS.gguf | IQ3_XS | 1.0GB |
| sarvam-2b-v0.5.IQ3_S.gguf | IQ3_S | 1.05GB |
| sarvam-2b-v0.5.Q3_K_S.gguf | Q3_K_S | 1.05GB |
| sarvam-2b-v0.5.IQ3_M.gguf | IQ3_M | 1.08GB |
| sarvam-2b-v0.5.Q3_K.gguf | Q3_K | 1.16GB |
| sarvam-2b-v0.5.Q3_K_M.gguf | Q3_K_M | 1.16GB |
| sarvam-2b-v0.5.Q3_K_L.gguf | Q3_K_L | 1.25GB |
| sarvam-2b-v0.5.IQ4_XS.gguf | IQ4_XS | 1.29GB |
| sarvam-2b-v0.5.Q4_0.gguf | Q4_0 | 1.35GB |
| sarvam-2b-v0.5.IQ4_NL.gguf | IQ4_NL | 1.36GB |
| sarvam-2b-v0.5.Q4_K_S.gguf | Q4_K_S | 1.36GB |
| sarvam-2b-v0.5.Q4_K.gguf | Q4_K | 1.43GB |
| sarvam-2b-v0.5.Q4_K_M.gguf | Q4_K_M | 1.43GB |
| sarvam-2b-v0.5.Q4_1.gguf | Q4_1 | 1.49GB |
| sarvam-2b-v0.5.Q5_0.gguf | Q5_0 | 1.62GB |
| sarvam-2b-v0.5.Q5_K_S.gguf | Q5_K_S | 1.62GB |
| sarvam-2b-v0.5.Q5_K.gguf | Q5_K | 1.67GB |
| sarvam-2b-v0.5.Q5_K_M.gguf | Q5_K_M | 1.67GB |
| sarvam-2b-v0.5.Q5_1.gguf | Q5_1 | 1.76GB |
| sarvam-2b-v0.5.Q6_K.gguf | Q6_K | 1.92GB |
| sarvam-2b-v0.5.Q8_0.gguf | Q8_0 | 2.48GB |
This is an early checkpoint of sarvam-2b, a small, yet powerful language model pre-trained from scratch on 4 trillion tokens. It is trained to be good at 10 Indic languages + English. Officially, the Indic languages supported are: Bengali, Gujarati, Hindi, Kannada, Malayalam, Marathi, Oriya, Punjabi, Tamil, and Telugu.
sarvam-2b will be trained on a data mixture containing equal parts English (2T) and Indic (2T) tokens. The current checkpoint has seen a total of 2 trillion tokens, and has not undergone any post-training.
Getting started:
from transformers import pipeline
pipe = pipeline(model='sarvamai/sarvam-2b-v0.5', device=0)
pipe('भारत के प्रथम प्रधानमंत्री', max_new_tokens=15, temperature=0.1, repetition_penalty=1.2)[0]['generated_text']
# 'भारत के प्रथम प्रधानमंत्री जवाहरलाल नेहरू की बेटी इंदिरा गांधी थीं।\n\n'
More technical details like evaluations and benchmarking will be posted soon.